Lecture 6

Lecture 6
Rohit Kumar
IIT Delhi
January 20, 2024
Rohit Kumar IIT Delhi January 20, 2024 1 / 51

Utility Maximization:
▶ Consumer choice problem can be formalized as
max u(x1 , x2 )
x1 ,x2
such that
p1 x1 + p2 x2 ≤ m
x1 , x2 ≥ 0.

Utility Assumptions:
utility function is continuous, increasing∗ and concave.
▶ We assumed that weak preferences are monotone.
▶ Here increasing utility means ∇u(x) ̸= 0 for all x.

Utility Maximization: Existence.
max u(x1 , x2 )
x1 ,x2
such that
p1 x1 + p2 x2 ≤ m
x1 , x2 ≥ 0.
▶ Existence:
1. utility function is continuous and
2. if prices are strictly positive, then budget set is compact.

Optimization Using the Lagrange Approach:
max u(x1 , x2 )
x1 ,x2
such that
p1 x1 + p2 x2 ≤ m
x1 , x2 ≥ 0.
▶ Can we convert the constrained maxima into unconstrained maxima?

▶ Consider the following minimization
min λ(m − p1 x1 − p2 x2 ).
λ≥0
▶ It is easy to see that

(
0 p1 x1 + p2 x2 ≤ m,
min λ(m − p1 x1 − p2 x2 ) =
λ≥0 −∞ p1 x1 + p2 x2 > m.

▶ Consumer choice problem can be rewritten as (u(0, 0) > −∞)
max u(x1 , x2 ) + min λ(m − p1 x1 − p2 x2 )

x1 ,x2 λ≥0
(
maxx1 ,x2 u(x1 , x2 ) p1 x1 + p2 x2 ≤ m,
=
−∞ p1 x1 + p2 x2 > m.

▶ Adding non negativity constraint as well (u(0, 0) > −∞)
max u(x1 , x2 ) + min λ(m − p1 x1 − p2 x2 ) + min µ1 (x1 ) + min µ2 (x2 )

x1 ,x2 λ≥0 µ1 ≥0 µ2 ≥0
(
maxx1 ,x2 u(x1 , x2 ) p1 x1 + p2 x2 ≤ m, x1 ≥ 0, x2 ≥ 0,
=
−∞ p1 x1 + p2 x2 > m or x1 < 0 or x2 < 0.

▶ The final optimization problem is
max min u(x1 , x2 ) + λ(m − p1 x1 − p2 x2 ) + µ1 x1 + µ2 x2 .

x1 ,x2 ∈R λ≥0,µ1 ≥0,µ2 ≥0

(a) First order condition for maximization
∂u(x∗1 , x∗2 )
− λ∗ p1 + µ∗1 = 0,
∂x1
∂u(x∗1 , x∗2 )
− λ∗ p2 + µ∗2 = 0.
∂x2
(b) Condition for minimization
λ∗ (m − p1 x∗1 − p2 x∗2 ) = 0, λ∗ ≥ 0, p1 x∗1 + p2 x∗2 ≤ m

µ∗1 x∗1 = 0, µ∗1 ≥ 0, x∗1 ≥ 0
µ∗2 x∗2 = 0, µ∗2 ≥ 0, x∗2 ≥ 0.

Necessary and Sufficient Condition:
Theorem
Suppose the utility function is continuously differentiable.
1. Any optimal solution satisfies the conditions (a)-(b).
2. If utility function is concave, then the solutions to (a)-(b) are optimal.

Necessary and Sufficient condition:
If utility function is concave, then FOC are necessary and sufficient.

Interior Solution:
Interior solution
x∗1 > 0, x∗2 > 0,
which implies
µ∗1 = 0, µ∗2 = 0.
because
µ∗1 x∗1 = 0, µ∗1 ≥ 0, x∗1 ≥ 0

µ∗2 x∗2 = 0, µ∗2 ≥ 0, x∗2 ≥ 0.

Interior Solution:
∂u(x∗1 , x∗2 )
− λ∗ p1 = 0,
∂x1
∂u(x∗1 , x∗2 )
− λ∗ p2 = 0.
∂x2
λ∗ (m − p1 x∗1 − p2 x∗2 ) = 0, λ∗ ≥ 0, p1 x∗1 + p2 x∗2 ≤ m

µ∗1 = 0, x∗1 > 0
µ∗2 = 0, x∗2 > 0.

Interior Solution: x∗1 > 0, x∗2 > 0, µ1 = 0, µ2 = 0.
▶ Case I: λ∗ > 0 and p1 x∗1 + p2 x∗2 = m.
▶ Suppose first order conditions are satisfied with x∗1 , x∗2 > 0.
▶ Then, we have
∂u(x∗ ∗
1 ,x2 )
∂x1 p1
∂u(x∗ ∗ = .
1 ,x2 ) p2
∂x2
▶ Marginal rate of substitution at x∗ is equal to p1 /p2
∂u(x∗ ∗
1 ,x2 )
∂x1 p1
−MRS(x∗ ) = − ∂u(x∗ ∗ =−
1 ,x2 ) p2
∂x2

x2
slope=− pp12
x∗
x1

▶ Case II: λ∗ = 0 and p1 x∗1 + p2 x∗2 = m.
▶ Then, we have
∂u(x∗1 , x∗2 ) ∂u(x∗1 , x∗2 )
= = 0.
∂x1 ∂x2
▶ This is impossible because u is differentiable and increasing*.

Concave and Increasing Function.
3
f (x)
x
5 10 15 20

Interior Solution: x∗1 > 0, x∗2 > 0.
▶ Case III: λ∗ = 0 and p1 x∗1 + p2 x∗2 < m.
▶ Then, we have
∂u(x∗1 , x∗2 ) ∂u(x∗1 , x∗2 )
= = 0.
∂x1 ∂x2
▶ This is impossible because u is differentiable and increasing*.

Interior Solution:
(a) MRS equals price ratio

∂u(x∗ ∗
1 ,x2 )
∂x1 p1
− ∂u(x∗ ∗ =− .
1 ,x2 ) p2
∂x2
(b) Point should lie on budget set
λ∗ > 0, p1 x∗1 + p2 x∗2 = m.

Cobb-Douglas Preferences:
▶ The optimization problem for Cobb-Douglas is (α, β ≥ 0 and α + β ≤ 1)

β
max min Axα
1 x2 + λ(m − p1 x1 − p2 x2 ) + µ1 x1 + µ2 x2 .
x1 ,x2 λ≥0,µ1 ≥0,µ2 ≥0
▶ We find the interior solution
µ∗1 = 0, µ∗2 = 0, λ∗ > 0, p1 x∗1 + p2 x∗2 = m.

▶ First order condition are (case I)
Aα(x∗1 )α−1 (x∗2 )β − λ∗ p1 = 0,

Aβ(x∗1 )α (x∗2 )β−1 − λ∗ p2 = 0,
m − p1 x∗1 − p2 x∗2 = 0, λ∗ > 0.
▶ Optimal solution x∗ satisfy
αx∗2 p1
= ,
βx∗1 p2
m = p1 x∗1 + p2 x∗2 .

▶ The optimal choice is

αm
x∗1 = ,
(α + β)p1
βm
x∗2 = .
(α + β)p2
▶ The expenditure share

αm
p1 x∗1 = ,
α+β
βm
p2 x∗2 = .
α+β

Leontief Preferences:
▶ The optimization problem for Leontief is
max min min{αx1 , βx2 } + λ(m − p1 x1 − p2 x2 ) + µ1 x1 + µ2 x2 .

x1 ,x2 λ≥0,µ1 ≥0,µ2 ≥0
▶ Lagrangian method only works when u is differentiable.
▶ Here, we will use graphical intuition to find the optimal solution.

x2
α
β
x∗
x1

▶ Optimal solution x∗ satisfy
x∗2 α
= ,
x∗1 β
m = p1 x∗1 + p2 x∗2 .
▶ The optimal choice is
βm
x∗1 = ,
p1 β + p2 α
αm
x∗2 = .
p1 β + p2 α

Corner Solution:
▶ Lets now focus on corner solution
either x∗1 = 0 or x∗2 = 0.
▶ x∗1 = 0 and x∗2 = 0 is not feasible as I can increase utility by consuming small
positive amount of both goods.

Corner Solution: Either x∗1 = 0 or x∗2 = 0.
∂u(x∗1 , x∗2 )
− λ∗ p1 + µ∗1 = 0,
∂x1
∂u(x∗1 , x∗2 )
− λ∗ p2 + µ∗2 = 0.
∂x2
λ∗ (m − p1 x∗1 − p2 x∗2 ) = 0, λ∗ ≥ 0, p1 x∗1 + p2 x∗2 ≤ m

µ∗1 x∗1 = 0, µ∗1 ≥ 0, x∗1 ≥ 0
µ∗2 x∗2 = 0, µ∗2 ≥ 0, x∗2 ≥ 0.

▶ For λ∗ , we have
λ∗ > 0, p1 x∗1 + p2 x∗2 = m.
▶ λ∗ is the shadow value of change in m.

▶ Case II: µ∗1 > 0, x∗1 = 0 and µ∗2 = 0, x∗2 > 0.
▶ This implies following optimal solution

m
x∗1 = 0, x∗ =
p2
and
∂u(x∗ ∗
1 ,x2 )
∂x1 + µ∗1 p1
∂u(x1 ,x∗
∗ = .
2) p2
∂x2

∂u(x∗ ∗
1 ,x2 )
∂x1 p1
−MRS(x∗ ) = − ∂u(x∗ ∗ >− .
1 ,x2 ) p2
∂x2

x2
x∗
x1

▶ Case III: µ∗1 = 0, x∗1 = 0 and µ∗2 = 0, x∗2 > 0.
▶ This implies following optimal solution

m
x∗1 = 0, x∗ =
p2
and
∂u(x∗ ∗
1 ,x2 )
∂x1 p1
∂u(x∗ ∗ = .
1 ,x2 ) p2
∂x2

∂u(x∗ ∗
1 ,x2 )
∂x1 p1
−MRS(x∗ ) = − ∂u(x∗ ∗ =− .
1 ,x2 ) p2
∂x2

x2
x∗
x1

▶ Case IV: µ∗1 = 0, x∗1 > 0 and µ∗2 > 0, x∗2 = 0.
∂u(x∗ ∗
1 ,x2 )
m ∂x1 p1
x∗1 = , x∗2 = 0, ∂u(x∗ ∗ = .
p1 1 ,x2 ) p2
∂x2 + µ∗2
▶ Case V: µ∗1 = 0, x∗1 > 0 and µ∗2 = 0, x∗2 = 0.
∂u(x∗ ∗
1 ,x2 )
m ∂x1 p1
x∗1 = , x∗2 = 0, ∂u(x∗ ∗ = .
p1 1 ,x2 ) p2
∂x2

x2 x2
Case IV Case V
x∗ x1 x∗ x1

▶ Case VI: µ∗1 ≥ 0, x∗1 = 0 and µ∗2 ≥ 0, x∗2 = 0.
▶ Not possible because solution must satisfy
p1 x∗1 + p2 x∗2 = m.

General Solution: Case I: Interior solution
∂u(x∗ ∗
1 ,x2 )
∂x1 p1 ∗
−MRS(x∗ ) = − ∂u(x∗ ∗ =− , λ > 0, p1 x∗1 + p2 x∗2 = m, µ∗1 = 0, µ∗2 = 0.
1 ,x2 ) p2
∂x2
x2
x∗
x1

General Solution: Case II: Corner solution x∗2 = 0
∂u(x∗ ∗
1 ,x2 )
∗ ∂x1 p1 ∗
−MRS(x ) = − ∂u(x∗ ∗ ≤− , λ > 0, p1 x∗1 + p2 x∗2 = m, µ∗1 = 0, µ∗2 = 0.
1 ,x2 ) p2
∂x2
x2 x2
x∗ x1 x∗ x1

General Solution: Case III: Corner solution x∗1 = 0.
∂u(x∗ ∗
1 ,x2 )
∗ ∂x1 p1 ∗
−MRS(x ) = − ∂u(x∗ ∗ ≥− , λ > 0, p1 x∗1 + p2 x∗2 = m, µ∗1 = 0, µ∗2 = 0.
1 ,x2 ) p2
∂x2
x2 x2
x∗ x∗
x1 x1

Linear Preference:
▶ The optimization problem for linear preference is
max min αx1 + βx2 + λ(m − p1 x1 − p2 x2 ) + µ1 x1 + µ2 x2

x1 ,x2 λ≥0,µ1 ≥0,µ2 ≥0
▶ Case I: interior solution.
▶ Case II: corner solution x∗2 = 0.
▶ Case III: corner solution x∗1 = 0.

Linear Preference:
x2
x1

Linear Preference:
▶ Case II: corner solution x∗2 = 0
x2
x∗ x1

Linear Preference:
▶ Case III: corner solution x∗1 = 0
x2
x∗
x1

Linear Preference:
▶ Optimal choice is

m α p1

 p1
 , 0 if β > p2
α p1
{(x1 , x2 ) : x1 ≥ 0, x2 ≥ 0, p1 x1 + p2 x2 = m} if β = p2

 0, m
 α p1
p2 if β < p2 .

Quasi-linear Preferences.
▶ The optimization problem for quasi-linear preference is
max min x1 + f (x2 ) + λ(m − p1 x1 − p2 x2 ) + µ1 x1 + µ2 x2

x1 ,x2 λ≥0,µ1 ≥0,µ2 ≥0
▶ Interior or corner solution depends on function f (a concave function).
▶ Lets take
√
f (x) = α x.

▶ Case I: interior solution.
2 x∗2
p
∗ p1
MRS(x ) = = , λ∗ > 0, p1 x∗1 + p2 x∗2 = m, µ∗1 = 0, µ∗2 = 0.
α p2
▶ Optimal solution
2
m − p2 x∗2

αp1
x∗2 = , x∗1 = .
2p2 p1

x2
x∗
x1

▶ General Solution.
( 2 )
αp1 m m − p2 x∗2
x∗2 = min , , x∗1 = .
2p2 p2 p1

▶ Corner solution. 2
αp1 m
≥ .
2p2 p2
x2 x2
x∗ x∗
x1 x1

Lecture 6

Uploaded by

Copyright:

Available Formats

Lecture 6

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture 6

Uploaded by

Copyright:

Available Formats

Lecture 6

January 20, 2024

Rohit Kumar IIT Delhi January 20, 2024 1 / 51

▶ Consumer choice problem can be formalized as

Rohit Kumar IIT Delhi January 20, 2024 2 / 51

utility function is continuous, increasing∗ and concave.

▶ We assumed that weak preferences are monotone.

▶ Here increasing utility means ∇u(x) ̸= 0 for all x.

Rohit Kumar IIT Delhi January 20, 2024 3 / 51

▶ Consumer choice problem can be formalized as

1. utility function is continuous and

2. if prices are strictly positive, then budget set is compact.

Rohit Kumar IIT Delhi January 20, 2024 4 / 51

▶ Consumer choice problem can be formalized as

▶ Can we convert the constrained maxima into unconstrained maxima?

Rohit Kumar IIT Delhi January 20, 2024 5 / 51

▶ Consider the following minimization

▶ It is easy to see that

Rohit Kumar IIT Delhi January 20, 2024 6 / 51

▶ Consumer choice problem can be rewritten as (u(0, 0) > −∞)

max u(x1 , x2 ) + min λ(m − p1 x1 − p2 x2 )

Rohit Kumar IIT Delhi January 20, 2024 7 / 51

▶ Adding non negativity constraint as well (u(0, 0) > −∞)

max u(x1 , x2 ) + min λ(m − p1 x1 − p2 x2 ) + min µ1 (x1 ) + min µ2 (x2 )

Rohit Kumar IIT Delhi January 20, 2024 8 / 51

▶ The final optimization problem is

max min u(x1 , x2 ) + λ(m − p1 x1 − p2 x2 ) + µ1 x1 + µ2 x2 .

Rohit Kumar IIT Delhi January 20, 2024 9 / 51

(a) First order condition for maximization

(b) Condition for minimization

λ∗ (m − p1 x∗1 − p2 x∗2 ) = 0, λ∗ ≥ 0, p1 x∗1 + p2 x∗2 ≤ m

Rohit Kumar IIT Delhi January 20, 2024 10 / 51

1. Any optimal solution satisfies the conditions (a)-(b).

2. If utility function is concave, then the solutions to (a)-(b) are optimal.

Rohit Kumar IIT Delhi January 20, 2024 11 / 51

If utility function is concave, then FOC are necessary and sufficient.

Rohit Kumar IIT Delhi January 20, 2024 12 / 51

x∗1 > 0, x∗2 > 0,

µ∗1 x∗1 = 0, µ∗1 ≥ 0, x∗1 ≥ 0

Rohit Kumar IIT Delhi January 20, 2024 13 / 51

(a) First order condition for maximization

(b) Condition for minimization

λ∗ (m − p1 x∗1 − p2 x∗2 ) = 0, λ∗ ≥ 0, p1 x∗1 + p2 x∗2 ≤ m

Rohit Kumar IIT Delhi January 20, 2024 14 / 51

▶ Case I: λ∗ > 0 and p1 x∗1 + p2 x∗2 = m.

▶ Marginal rate of substitution at x∗ is equal to p1 /p2

Rohit Kumar IIT Delhi January 20, 2024 15 / 51

Rohit Kumar IIT Delhi January 20, 2024 16 / 51

▶ Case II: λ∗ = 0 and p1 x∗1 + p2 x∗2 = m.

▶ This is impossible because u is differentiable and increasing*.

Rohit Kumar IIT Delhi January 20, 2024 17 / 51

Rohit Kumar IIT Delhi January 20, 2024 18 / 51

▶ Case III: λ∗ = 0 and p1 x∗1 + p2 x∗2 < m.

▶ This is impossible because u is differentiable and increasing*.

Rohit Kumar IIT Delhi January 20, 2024 19 / 51

(a) MRS equals price ratio

(b) Point should lie on budget set

λ∗ > 0, p1 x∗1 + p2 x∗2 = m.